As smart and self-driving vehicles gain more momentum, the technologies around detection systems improve and become more reliable. And as humans who occasionally cross streets, who wouldn’t want improved pedestrian detection?First picked up by Phys.org, engineers at the University of California, San Diego have been working on the CompACT detection system for vehicles, which specifically looks for pedestrians. According to the team’s report, the CompACT detectors film and analyze at 2-4 frames per second, detecting people in real time based on a complicated algorithm, and it’s pretty accurate.SEE ALSO: I tested Tesla Autopilot in Manhattan traffic — and lived to tell about itPedestrian detection isn’t anything new in cars — Volvo first introduced detection in 2010, which recognized vehicles and people as separate objects using a front-facing camera and radar. Volvo has also worked on detection technology for cyclists and animals.Some newer vehicle models include detection and response systems, like the 2017 Mercedes-Benz E-Class, which detects pedestrians and calculates what it needs to do with the car’s steering and brakes to avoid them. But detection has room to improve, as false-positives and non-detections can lead to some avoidable, dangerous situations.The CompACT improves upon standard detection by combining normal cascading detection with a more complex algorithm. Cascading detection works by ruling out sections of a video feed that are definitely not pedestrians, like the road or the sky, and moving deeper into smaller and more detailed sections of the image to find people. Most pedestrian detectors don’t get more detailed than that, according to the report, but the CompACT’s final process is much more accurate.